CN114550019A - Edge end in-place monitoring system based on target detection algorithm - Google Patents
Edge end in-place monitoring system based on target detection algorithm Download PDFInfo
- Publication number
- CN114550019A CN114550019A CN202210346966.5A CN202210346966A CN114550019A CN 114550019 A CN114550019 A CN 114550019A CN 202210346966 A CN202210346966 A CN 202210346966A CN 114550019 A CN114550019 A CN 114550019A
- Authority
- CN
- China
- Prior art keywords
- detection
- image
- target detection
- edge end
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/60—Electric or hybrid propulsion means for production processes
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an edge end in-place monitoring system based on a target detection algorithm, which belongs to the technical field of logistics equipment and comprises an intelligent in-place detection module and a management platform; the intelligent in-place detection module is arranged on the AGV and comprises edge end equipment, an image acquisition unit and a power supply unit; the edge terminal equipment comprises an image acquisition unit, an image detection unit and a cargo identification unit; according to the invention, a YOLOv3 and SSD target detection algorithm are adopted to carry out sample training to obtain a complete edge end target detection model, compared with the traditional two-stage method, the detection and recognition speed is higher, and the model is continuously optimized through repeated testing and sample addition, so that the detection and recognition accuracy of the model is improved; in addition, the detection and the identification of the product image are realized on the edge terminal equipment, so that the data processing pressure of the management platform is reduced.
Description
Technical Field
The invention relates to the technical field of logistics equipment, in particular to an edge end on-position monitoring system based on a target detection algorithm.
Background
Through retrieval, the Chinese patent No. CN108750509A discloses a goods position detection method based on an AGV and the AGV, the method detects the current condition of the goods position through a sensor, although the misoperation of the AGV in the process of taking and placing materials is reduced, the method only can detect whether goods exist on the goods position and can not detect the category of the goods, and further can not assist a goods management platform to carry out real-time goods monitoring and management; an Automated Guided Vehicle (AGV) trolley is a transport Vehicle equipped with an automatic guide device, capable of automatically traveling along a specified guide path and having multiple automatic transfer functions, and at present, along with the development of electronic technology, the application range of the AGV is continuously expanded, and the AGV trolley is most widely applied in the aspect of logistics systems, but the function of the existing AGV trolley is relatively single, and only the logistics transportation function can be realized; in an actual logistics warehouse, due to the fact that the number of goods is large and the goods are frequently circulated, the product conditions of each storage position of a goods shelf need to be known in real time; however, most of the existing methods are realized in a manual inspection mode, so that the time and labor are wasted, and errors are easy to occur; therefore, how to combine the AGV and the target detection system to be applied to the detection and identification of the bin becomes an important research topic at present; therefore, it becomes more important to invent an edge end on-site monitoring system based on a target detection algorithm;
the existing in-place monitoring system can only detect whether goods exist on a goods position and can not detect the type of the goods, so that the existing in-place monitoring system can not assist a goods management platform to carry out real-time goods monitoring management; for this purpose, we propose an edge-end on-site monitoring system based on an object detection algorithm.
Disclosure of Invention
The invention aims to solve the defects in the prior art, and provides an edge end on-position monitoring system based on an object detection algorithm.
In order to achieve the purpose, the invention adopts the following technical scheme:
the edge end in-place monitoring system based on the target detection algorithm comprises an intelligent in-place detection module and a management platform;
the intelligent in-place detection module is arranged on the AGV and comprises edge end equipment, an image acquisition unit and a power supply unit; the edge terminal equipment comprises an image acquisition unit, an image detection unit and a cargo identification unit; the intelligent in-place detection module and the management platform are in remote communication through a wireless communication technology.
Furthermore, the image acquisition unit is specifically a camera and is used for shooting images of all bins on the goods shelf in the running process of the AGV trolley and acquiring product images; the power supply unit is used for supplying power to the intelligent in-place detection module.
Further, the image acquisition unit is used for acquiring the product image acquired by the image acquisition unit; the image detection unit is used for detecting a product image through a complete edge end target detection model to obtain an image detection result, and the image detection result comprises two conditions: one is that the position has goods, the other is that the position has no goods; the goods identification unit is used for carrying out category identification on the product image with goods in the image detection result through the complete edge end target detection model to obtain a goods category identification result.
Further, the specific forming process of the complete edge end target detection model is as follows:
s1: firstly, collecting various product images through a detection unit, and carrying out category marking on the product images to form a required sample data set;
s2: secondly, performing target detection algorithm training on the sample data set in the step S1 by adopting a MobileNet lightweight target detection network to form an initial edge end detection model I;
s3: then, deploying the initial edge detection model in the step S2 to an edge device through hardware, and running the initial edge detection model in all actually occurring scenes to test the model performance;
s4: acquiring product images identified by all scenes in the testing process, and carrying out negative sample labeling on the product images failed in detection and identification to obtain a failed product image sample set;
s5: adding the failed product image sample set obtained in the step S4 into the sample data set obtained in the step S1 to form a new sample data set, and performing target detection algorithm training on the new sample data set by adopting a MobileNet lightweight target detection network again;
s6: then, cutting, quantifying and distilling the target detection model in the step S5 to form an initial edge end detection model II;
s7: and repeating the steps S3-S6 until the performance of the model reaches the optimum, and obtaining the complete edge end target detection model.
Further, the annotated formats include, but are not limited to, VOC formats and COCO formats; the hardware includes but is not limited to Jeston Nano, RK3399Pro development board, and raspberry pie; the target detection algorithm is specifically YOLOv3 and SSD.
Further, the wireless communication technologies include, but are not limited to, WIFI, 4G, and 5G; and the management platform is used for receiving the image detection result and the goods category identification result and carrying out service processing according to the image detection result and the goods category identification result.
Compared with the prior art, the invention has the beneficial effects that:
1. the edge end on-site monitoring system based on the target detection algorithm performs sample training by adopting the YOLOv3 and SSD target detection algorithm to obtain a complete edge end target detection model, compared with the traditional two-stage method, the detection and identification speed is higher, and the model is continuously optimized by repeated testing and sample addition, so that the detection and identification accuracy of the model is improved; in addition, the detection and the identification of the product image are realized on the edge terminal equipment, so that the data processing pressure of a management platform is reduced;
2. according to the edge end in-place monitoring system based on the target detection algorithm, a complete edge end target detection model is combined with an AGV through hardware with small size (such as Jeston Nano, RK3399Pro development board and raspberry group), so that the system is convenient to carry, and can be applied to multiple fields of industrial production and transportation.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention.
FIG. 1 is a schematic diagram of an overall structure of an edge end on-site monitoring system based on a target detection algorithm according to the present invention;
FIG. 2 is a schematic structural diagram of an AGV trolley collocated with an edge end in-place monitoring system based on a target detection algorithm provided by the present invention;
fig. 3 is a schematic flow chart of a complete edge end target detection model forming process in the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments.
In the description of the present invention, it is to be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", and the like, are used in the orientations and positional relationships indicated in the drawings, which are based on the orientations and positional relationships indicated in the drawings, and are used for convenience of description and simplicity of description, but do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, are not to be construed as limiting the present invention.
Referring to fig. 1-2, the embodiment discloses an edge end in-place monitoring system based on a target detection algorithm, which comprises an intelligent in-place detection module and a management platform;
the intelligent in-place detection module is arranged on the AGV trolley and comprises edge end equipment, an image acquisition unit and a power supply unit; the edge terminal equipment comprises an image acquisition unit, an image detection unit and a cargo identification unit; the intelligent in-place detection module and the management platform are in remote communication through a wireless communication technology.
The image acquisition unit is specifically a camera and is used for shooting images of all bins on the goods shelf in the running process of the AGV trolley and acquiring product images;
specifically, the image acquisition unit needs to adjust the camera to the optimal position according to different scene installation conditions, so as to realize the full coverage of the product visual angle;
the power supply unit is used for supplying power to the intelligent in-place detection module.
The image acquisition unit is used for acquiring the product image acquired by the image acquisition unit; the image detection unit is used for detecting the product image through the complete edge end target detection model to obtain an image detection result, and the image detection result comprises two conditions: one is that the position has goods, the other is that the position has no goods;
specifically, the image detection result without goods is directly reported to the management platform, and the image detection result with goods is uploaded to the management platform after category identification is carried out on the image detection result with goods;
the goods identification unit is used for carrying out category identification on the product image with goods in the image detection result through the complete edge end target detection model to obtain a goods category identification result.
Wireless communication technologies include, but are not limited to, WIFI, 4G, and 5G;
specifically, the wireless communication technology may be selected according to the actual communication distance and the expected communication cost, and is not limited to those listed in this example;
and the management platform is used for receiving the image detection result and the goods category identification result and carrying out service processing according to the image detection result and the goods category identification result.
Referring to fig. 3, the embodiment discloses an edge end in-place monitoring system based on a target detection algorithm, which includes an intelligent in-place detection module and a management platform;
the intelligent in-place detection module is arranged on the AGV and comprises edge end equipment, an image acquisition unit and a power supply unit; the edge terminal equipment comprises an image acquisition unit, an image detection unit and a cargo identification unit; the intelligent in-place detection module and the management platform are in remote communication through a wireless communication technology.
Except for the same structure as the above embodiment, the present embodiment will specifically describe the formation process of the complete edge object detection model;
specifically, the complete edge end target detection model is formed in the following specific process:
firstly, collecting various product images through a detection unit, and carrying out category marking on the product images to form a required sample data set; secondly, performing target detection algorithm training on the sample data set by adopting a MobileNet lightweight target detection network to form a first initial edge end detection model; then, deploying the initial edge detection model to edge equipment through hardware, and running the initial edge detection model in all actually occurring scenes to test the performance of the model; acquiring product images identified by all scenes in the testing process, and carrying out negative sample labeling on the product images failed in detection and identification to obtain a failed product image sample set; adding the failed product image sample set into the sample data set to form a new sample data set, and performing target detection algorithm training on the new sample data set by adopting a MobileNet lightweight target detection network again; then, cutting, quantifying and distilling the target detection model to form an initial edge end detection model II; and repeating the steps until the performance of the model reaches the optimum, thus obtaining the complete edge end target detection model.
Specifically, the tagged formats include, but are not limited to, VOC format and COCO format; the hardware includes but is not limited to Jeston Nano, RK3399Pro development board, and raspberry pie; the target detection algorithm is embodied as YOLOv3 and SSD.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.
Claims (6)
1. The edge end in-place monitoring system based on the target detection algorithm is characterized by comprising an intelligent in-place detection module and a management platform;
the intelligent in-place detection module is arranged on the AGV and comprises edge end equipment, an image acquisition unit and a power supply unit; the edge terminal equipment comprises an image acquisition unit, an image detection unit and a cargo identification unit; the intelligent in-place detection module and the management platform are in remote communication through a wireless communication technology.
2. The target detection algorithm-based edge end in-place monitoring system according to claim 1, wherein the image acquisition unit is specifically a camera, and is used for shooting images of various positions on a shelf in the running process of the AGV trolley and acquiring product images; the power supply unit is used for supplying power to the intelligent in-place detection module.
3. The target detection algorithm-based edge end in-place monitoring system of claim 1, wherein the image acquisition unit is configured to acquire the product image acquired by the image acquisition unit; the image detection unit is used for detecting a product image through the complete edge end target detection model to obtain an image detection result, and the image detection result comprises two conditions: one is that the position has goods, the other is that the position has no goods; the goods identification unit is used for carrying out category identification on the product image with goods in the image detection result through the complete edge end target detection model to obtain a goods category identification result.
4. An edge end on-site monitoring system based on an object detection algorithm according to claim 3, wherein the complete edge end object detection model is formed in the following specific process:
s1: firstly, collecting various product images through a detection unit, and carrying out category marking on the product images to form a required sample data set;
s2: secondly, performing target detection algorithm training on the sample data set in the step S1 by adopting a MobileNet lightweight target detection network to form an initial edge end detection model I;
s3: then, deploying the initial edge detection model in the step S2 to an edge device through hardware, and running the initial edge detection model in all actually occurring scenes to test the model performance;
s4: acquiring product images identified by all scenes in the testing process, and carrying out negative sample labeling on the product images failed in detection and identification to obtain a failed product image sample set;
s5: adding the failed product image sample set obtained in the step S4 into the sample data set obtained in the step S1 to form a new sample data set, and performing target detection algorithm training on the new sample data set by adopting a MobileNet lightweight target detection network again;
s6: then, cutting, quantifying and distilling the target detection model in the step S5 to form an initial edge end detection model II;
s7: and repeating the steps S3-S6 until the performance of the model reaches the optimum, and obtaining the complete edge end target detection model.
5. An edge-end on-site monitoring system based on object detection algorithm as claimed in claim 4, characterized in that the tagged format includes but is not limited to VOC format and COCO format; the hardware includes but is not limited to Jeston Nano, RK3399Pro development board, and raspberry pie; the target detection algorithm is specifically YOLOv3 and SSD.
6. An edge-end on-site monitoring system based on an object detection algorithm as claimed in claim 1, wherein the wireless communication technologies include but are not limited to WIFI, 4G and 5G; and the management platform is used for receiving the image detection result and the goods category identification result and carrying out service processing according to the image detection result and the goods category identification result.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110744148.6A CN113326811A (en) | 2021-07-01 | 2021-07-01 | Edge end in-place monitoring system based on target detection algorithm |
| CN2021107441486 | 2021-07-01 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114550019A true CN114550019A (en) | 2022-05-27 |
Family
ID=77425434
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110744148.6A Withdrawn CN113326811A (en) | 2021-07-01 | 2021-07-01 | Edge end in-place monitoring system based on target detection algorithm |
| CN202210346966.5A Withdrawn CN114550019A (en) | 2021-07-01 | 2022-04-01 | Edge end in-place monitoring system based on target detection algorithm |
Family Applications Before (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110744148.6A Withdrawn CN113326811A (en) | 2021-07-01 | 2021-07-01 | Edge end in-place monitoring system based on target detection algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (2) | CN113326811A (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116524017B (en) * | 2023-03-13 | 2023-09-19 | 明创慧远科技集团有限公司 | Underground detection, identification and positioning system for mine |
-
2021
- 2021-07-01 CN CN202110744148.6A patent/CN113326811A/en not_active Withdrawn
-
2022
- 2022-04-01 CN CN202210346966.5A patent/CN114550019A/en not_active Withdrawn
Also Published As
| Publication number | Publication date |
|---|---|
| CN113326811A (en) | 2021-08-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108189043B (en) | Inspection method and inspection robot system applied to high-speed rail machine room | |
| CN112525917A (en) | Battery cell tab detection method, device and system and storage medium | |
| CN114550019A (en) | Edge end in-place monitoring system based on target detection algorithm | |
| CN112651516A (en) | Intelligent inspection system and method for transformer substation machine room based on AR technology | |
| CN110045704A (en) | The retroactive method of creation data | |
| CN114399258A (en) | Intelligent goods shelf, warehousing system based on intelligent goods shelf and management method thereof | |
| CN111274934A (en) | Implementation method and system for intelligently monitoring forklift operation track in warehousing management | |
| CN112087528B (en) | Intelligent water environment monitoring system and method based on deep learning | |
| CN109975304B (en) | Automatic circulation full-automatic fault indicator assembly line detection system and method | |
| CN208536899U (en) | A kind of Cold Chain Logistics vehicle monitoring system based on big data | |
| CN109975728B (en) | Intelligent full-automatic streamline comprehensive detection system and method for fault indicator | |
| CN115880231A (en) | Power transmission line hidden danger detection method and system based on deep learning | |
| CN111376789A (en) | AGV battery charging management method, system and computer readable storage medium | |
| CN109242843B (en) | Container surface image and container number image acquisition method | |
| CN207069695U (en) | A kind of transformer station's O＆M quantifies cruising inspection system | |
| CN112540570A (en) | Automatic inspection management method and system for file warehouse robot | |
| CN109975729B (en) | Full-automatic power distribution automation terminal comprehensive detection system and method | |
| CN113055658A (en) | Tunnel hazardous chemical substance vehicle identification method and system based on panoramic stitching technology | |
| CN206517529U (en) | A kind of unmanned aerial vehicle platform camera system | |
| CN113359829A (en) | Unmanned aerial vehicle power plant intelligent inspection method based on big data | |
| CN112164095A (en) | ROS-based airborne target detection and tracking method | |
| CN204613128U (en) | Circuit board detection device | |
| CN113721605B (en) | Cylinder jacket goods shelf placement self-identification system | |
| CN215987376U (en) | Photovoltaic module hot spot unmanned aerial vehicle automatic check out system based on two optical technologies | |
| CN209486197U (en) | A kind of transformer substation robot cruising inspection system of combination account |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20220527 |